Localising the CEFR Material: Teachers’ Perceptions of the C.R.A.F.T. Guide for AI-Assisted Adaptation

The adoption of the Common European Framework of Reference (CEFR) in Malaysia has undoubtedly exacerbated the cultural divide for rural ESL learners, who frequently encounter difficulties in reading comprehension due to unfamiliar foreign schemata. Although Generative AI offers a means for material adaptation, the absence of validated rapid engineering methods impedes its practical implementation in educational settings. This study presents the C.R.A.F.T. (Culturally Responsive AI Framework for Teachers) Guide, a systematic procedure designed to democratise the development of Schema-aligned resources. This research employs an exploratory multiple case study methodology to evaluate the usability and effectiveness of the framework with three secondary school teachers from Johor, Sarawak, and Pulau Pinang. Data were triangulated through non-participant observations and semi-structured interviews. The results demonstrate that the C.R.A.F.T. Guide diminished material adaptation processes to less than 15 minutes per text and significantly transformed the teacher's role from 'content developer' to 'cultural gatekeeper.' The results indicate that structured AI frameworks can effectively connect standardised competency goals with local cultural authenticity.